It's your choice! New Modular Organization! Appncatirms emphasis: A course that covers the principles of database systems and emphasizes how they are used in developing data-intensive applications f,;~tY'W';Yl~t';;:;,~7' A course that has a strong systems emphasis and assumes that students have good programming skills in C and C++ Hybrid course: Modular organization allows you to teach the course with the emphasis you want - := I ~~~ ER Model Conceptual Design II III IV v I VIr 27 Infonnation Retrieval and XML Data Management Relational Model SQLDDL Dependencies j j j j j j j j j j j j j j j j j j j j j j j j j j j j j j DATABASE MANAGEMENT SYSTEMS DATABASE MANAGEMENT SYSTEMS Third Edition Raghu Ramakrishnan University of Wisconsin Madison, Wisconsin, USA • Johannes Gehrke Cornell University Ithaca, New York, USA Boston Burr Ridge, IL Dubuque, IA Madison, WI New York San Francisco St Louis Bangkok Bogota Caracas Kuala Lumpur Lisbon London Madrid Mexico City Milan Montreal New Delhi Santiago Seoul Singapore Sydney Taipei Toronto McGraw-Hill Higher Education tz A Lhvision of The McGraw-Hill Companies DATABASE MANAGEMENT SYSTEMS, THIRD EDITION International Edition 2003 Exclusive rights by McGraw-Hill Education (Asia), for manufacture and export This book cannot be re-exported from the country to which it is sold by McGraw-Hill The International Edition is not available in North America Published by McGraw-Hili, a business unit of The McGraw-Hili Companies, Inc., 1221 Avenue of the Americas, New York, NY 10020 Copyright © 2003, 2000, 1998 by The McGraw-Hill Companies, Inc All rights reserved No part of this publication may be reproduced or distributed in any form or by any means, or stored in a database or retrieval system, without the prior written consent of The McGraw-Hill Companies, Inc., including, but not limited to, in any network or other electronic storage or transmission, or broadcast for distance learning Some ancillaries, including electronic and print components, may not be available to customers outside the United States 10 09 08 07 06 05 04 03 20 09 08 07 06 05 04 CTF BJE Library of Congress Cataloging-in-Publication Data Ramakrishnan, Raghu Database management systems / Raghu Ramakrishnan, Johannes p Gehrke.~3rd cm Includes index ISBN 0-07-246563-8-ISBN 0-07-115110-9 (ISE) Database management Gehrke, Johannes II Title QA76.9.D3 R237 2003 005.74 Dc21 When ordering this title, use ISBN 0-07-123151-X Printed in Singapore www.mhhe.com 2002075205 CIP ed To Apu, Ketan, and Vivek with love To Keiko and Elisa 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 CONTENTS PREFACE Part I XXIV FOUNDATIONS OVERVIEW OF DATABASE SYSTEMS 1.1 1.2 Managing Data A Historical Perspective 1.3 File Systems versus a DBMS 1.4 1.5 Advantages of a DBMS Describing and Storing Data in a DBMS 1.5.1 The Relational Model 1.5.2 1.6 Levels of Abstraction in a DBMS 10 11 12 15 1.5.3 Data Independence Queries in a DBMS Transaction Management 1.7.1 Concurrent Execution of Transactions 1.7.2 1.7.3 Incomplete Transactions and System Crashes Points to Note 16 17 17 18 1.8 1.9 1.10 Structure of a DBMS 19 19 People Who Work with Databases Review Questions 21 22 INTRODUCTION TO DATABASE DESIGN 2.1 Database Design and ER Diagrams 2.1.1 2.2 2.3 2.4 Beyond ER Design Entities, Attributes, and Entity Sets Relationships and Relationship Sets Additional Features of the ER Model 25 26 27 28 29 32 2.4.1 2.4.2 Key Constraints Participation Constraints 32 34 2.4.3 Weak Entities 2.4.4 2.4.5 Class Hierarchies Aggregation 35 37 39 vii Ii [fTH()R LNDE.:X; 'Nang: H., 888 l 102:3 Ward, K., 516, 1021 \Varren, D.S., 844-845, 1030, 10:35, 10:37, 1041 \Vatson, V., 98, 1007 \Neber, R., 991, 1043 \Veddell, G.E., 648, 1043 \Vei, J., 10~~9 \Veihl, \V., 602, 1043 \Veikum, G., 3:37, 548, 815, 1034, 1037, 1043 Weiner, J., 967, 1032 \Veinreb, D., 815, 1028 \Veiss, R., 966, 1043 Wenger, K., 1029, 1001 West, IVI., 520 Whitaker, M., xxxii White, C., 771, 104:3 White, S., 219, 1043 White, S.J., 815, 1011 Widom, J., 24, 99, 181, 771, 887 -888, 967, 1006-1007, 1012, 1020 ·1021, 1030, 1033·1034, 1043~1044 Wiederhold, G., 24, xxix, 30:3, :337, 771, 887, 1020, 1031, 1034, 1043 Wilkinson, W.K., 438, 477, 578, 1008, 1027 Willett, P., 966, 1025 Williams, R., 771, 1043 Wilms, P.P., 771, 815, 1022, 1043 Wilson, L.O., 924, 1038 \Vinuner, 1'v1., 925 \Vimmers, E.L., 925, 1006 \Vinslett, M.S., 99, 722, 10:39, 104:3 \iViorkowski, G., 691, 1043 \Vise, T.E., :3:n, 1008 \Vistrand, E., 1006, 1001 \Vitten, LH., 924, 966, 104:3 Woelk, D., 815, 1026 vVolfson, 0., 771, 991, 1024, 1043, 1001 \Yong, C.Y., 925, 1014 vVong, E., 516, 771, 1009, 1017, 1036, 1043 Wong, H.K.T., 516, 1020 Wong, L., 816, 1011 'Nang, W., 548, 1009 Wood, D., 477, 1016 Woodruff, A., 1006, 1001 Wright, F.L., 649 Wu, J., 816, 1026 Wylie, K., 888, 1007 Xu, E., 991, 1043 Xu, X., 925, 1017, 1037 Yajima, S., 771, 1025 Yang, D., 56, 99, 1041 Yang, Y., 924, 1043 Yannakakis, Iv!., 516, 1036 Yao, S.B., 578, 1028 Yin, Y., 924, 1023 Yoshikawa, Iv!., 771, 816, 1024 -1025 Yossi, 11., 887 -888 Yost, R.A., 98, 771, 1012, 104:3 Young, H.C., 4:38, 1029 Youssefi, K., 516, 104:3 Yuan, L., 816, 10:3:3 Ya, C.T., 771, 104:31044 '{u, J-B., m.n, 1021, 10:34, 1001 Vue, K.B., xxxi Yurttas, S.: xx :xi Zaiane, O.R., 924, 104~3 Zaki, IvLJ., 924, 1044 Zaniolo, C., 98, 181: 516, 648, 816, 844 · 845, 1014, 1027, 1036, 1044, 1001 Zaot, :M., 925, 1006 Zdonik, S.B., x,xix, 516, 816, 888, 925, 1031, 1038, 1040, 1044 Zhang, A., 771, 1017 Zhang, T., 925, 1044 Zhang, VV., 1044 Zhao, W., 1040, 1001 Zhao, Y., 887, 1044 Zhou, J., 991, 1043 Zhuge, Y., 887, 1044 Ziauddin, M., xxxi Zicari, R., 181, 816, 844, 1044, 1001 Zilio, D.C., 691, 1042 Zloof, Iv1.M., x,xix, 98, 1044 Zobel, J., 966, 1031, 1044 Zukowski, U., 844, 1044 Zuliani, ~1., 691, 1042 Zwilling, M.J., 815, 1011 SUBJECT INDEX INF, 615 2NF, 619 2PC, 759, 761 blocking, 760 with Presumed Abort, 762 2PL, 552 distributed databases, 755 3NF, 617, 625, 628 3PC, 762 4NF, 636 5NF, 638 A priori property, 893 Abandoned privilege, 700 Abort, 522 523, 533, 535, 583, 593, 759 Abstract data types, 784 785 ACA schedule, 530 Access control, 9, 693-694 Access invariance, 569 Access mode in SQL, 538 Access path, 398 most selective, 400 Access privileges, 695 Access times for disks, 284, 308 ACID transactions, 521 Active databases, 132, 168 Adding tables in SQL, 91 Adorned program, 839 ADTs, 784 785 encapsulation, 785 storage issues, 799 Advanced Encryption Standard (AES), 710 AES, 710 Aggregate functions in ORDBMSs, 801 Aggregation in Datalog, 8:31 Aggregation in SQL, 151, 164 Aggregation in the ER model, :39,84 Algebra relational, 102 ALTER, 696 Alternatives for data entries in an index, 276 Analysis phase of recovery, 580, 588 ANSI, 6, 58 API, 195 Applica.tion architectures, 2:36 Application programmers, 21 Application programming interface, 195 Application servers, 251, 253 Architecture of a DBIvIS, 19 ARIES recovery algorithm, 543, 580, 596 Armstrong's Axioms, 612 Array chunks, 800, 870 Arrays, 781 Assertions in SQL, 167 Association rules, 897, 900 use for prediction, 902 with calendars, 900 with item hierarchies, 899 Asynchronous replication, 741, 750 751, 871 Capture and Apply, 752 -753 change data table (CDT), 753 conflict resolution, 751 peer-to-peer, 751 primary site, 751 Atomic formulas, 118 Atomicity, 521-522 Attribute, 11 Attribute closure, 614 Attributes in the ER model, 29 Attributes in the relational model, 59 Attributes in XNIL, 229 Audit trail, 715 Authentication, 694 Authorities, 941 Authorization, 9, 22 Authorization graph, 701 Authorization ID, 697 Autocommit in JDBC, 198 AVC set, 909 AVG, 151 Avoiding cascading aborts, 5;~0 Axioms for FDs, 612 B+ trees, 281, 344 bulk-loading, :360 deletion, :352 for sorting, 4:33 height, ;3115 insertion, :.~48 key compression, :358 locking, 561 order, ;~45 1054 search, 347 selection operation, 442 sequence set, 345 B+ trees vs ISA~1, 292 Bags, 780, 782 Base table, 87 BCNF, 616, 622 Bell-LaPadula security model, 706 Benchmarks, 506, 683, 691 Binding early vs late, 788 Bioinformatics, 999 BIRCH, 912 Birth site, 742 Bit-sliced signature files, 939 Bitmap indexes, 866 Bitmapped join index, 869 Bitmaps for space management, 317, 328 Blind writes, 528 BLOBs, 775, 799 Block evolution of data, 916 Block nested loops join, 455 Blocked I/O, 430 Blocking, 5:3:3, 865 Blocks in disks, 306 Bloomjoin, 748 Boolean queries, 929 Bounding box, 982 Boyce-Codd nonnal form, 616, 622 Buckets, 279 in a hashed file, :371 in histograms, 486 Buffer frame, 318 Buffer management DBMS VS OS, 322 double bufl'ering, 432 force approach, 541 real systems, :322 replacernent policy, :321 sequential flooding, ;321 steal approach, 541 Buffer m,luager, 20, :305, :318 forcing a page, :323 page replacement, :ng·320 pinning, ;U9 prefetching, :322 SUBJECT INDEX Buffer pool, ~n8 Buffered writes, 571 Building phase in hash join, 46:3 Bulk data types, 780 Bulk-loading 13+ trees, 360 Bushy trees, 415 Caching of methods, 802 CAD jCA:M, 971 Calculus relational, 116 Calendric a ssociation rules, 900 Candidate keys, 29, 64, 76 Capture and Apply, 752 Cardinality of a relation, 61 Cartsian product, 105 CASCADE in foreign keys, 71 Cascading aborts, 530 Cascading operators, 488 Cascading Style Sheets, 249 Catalogs, 394 395, 480, 483, 741 Categorical attribute, 905 Centralized deadlock detection, 756 Centralized lock management, 755 Certification authorities, 712 CGI, 251 Chained transactions, 536 Change data table, 753 Change detection, 916· 917 Character large object, 776 Checkpoint, 19, 587 fuzzy, 587 Checkpoints, 543 Checksum, 307 Choice of indexes, 653 Chunking, 800, 870 Class hierarchies, :37, 8:3 Class interface, 806 Classification, 904-905 Classification rules, 905 Cla.'ssification trees, 906 Clearance, 706 Client-server architecture, 2:37, 738 CL013, 776 Clock, 322 Clock policy, ~321 Close an iterator, 408 Closure of 1"Ds, 612 CLR s, 584, 592, 596 Clustered file, 277 Clustered files, 287 Clustering, 277, 29:~, 660, 911 CODASYL, D.B.T.G., 1014 Collations in SQL, 140 Collection hierarchies, 789 Collection hierarchy, 789 1055 ~ Collection types, 780 Collisions, :379 Column, 59 Commit, 523, 535, 58:3, 759 Commit protocols, 751, 758 2PC, 759, 761 3PC,762 Communication costs, 7:39, 744, 749 Communication protocol, 223 Compensation log records, 584, 592, 596 Complete axioms, 613 Complex types, 779, 795 vs reference types, 795 Composite search keys, 295, 297 Compressed histogram, 487 Compression in B+ trees, 358 Computer aided design and manufacturing, 971 Concatenated search keys, 295, 297 Conceptual design, 13, 27 tuning, 669 Conceptual evaluation strategy, 133 Conceptual schema, 13 Concurrency, 9, 17 Concurrency control multiversion, 572 optimistic, 566 timestamp, 569 Concurrent execution, 524 Conflict equivalence, 550 Conflict resolution, 751 Conflict serializability vs serializability, 561 Conflict serializable schedule, 550 Conflicting actions, 526 Conjunct, 445 primary, :399 Conjunctive normal form (CNF), :398, 445 Connection pooling, 200 Connections in IDBC, 198 Conservative 2PL, 559 Consistency, 521 Content types in X1vlL, 2:32 Content-ba",sed queries, 972, 988 Convoy phenomenon, 555 Cookie, 259 Cookies, 2,5~3 Coordinator site, 758 Correlated queries, 147, 504, 506 Cosine normalization, 9:32 Cost estirnatioIl, 48248:3 for ADT methods 803 real systems, 485 Cost model, 440 COUNT, 151 Covering constraints, 38 Covert channel, 708 Crabbing, 5fi2 Crash recovery, 9, 18, 22, 541, 580, 583~584, 587 588, 590, 592, 595-596 Crawler, 9:39 CREATE DOfvlAIN, 166 CREATE statement SQL, 696 CREATE TABLE, 62 CREATE TRIGGER, 169 CREATE TYPE, 167 CREATE VIEW, 86 Creating a relation in SQL, 62 Critical section, 567 Cross-product operation, 105 Cross-tabulation, 855 C8564 at Wisconsin, xxviii CSS,249 CUBE operator, 857, 869, 887 Cursors in SQL, 189, 191 Cylinders in disks, 306 Dali, 1001 Data definition language, 12 Data Definition Language (DDL), 12, 62, 131 Data dictionary, 395 Data Encryption Standard, 710 Data Encryption Standard (DES), 710 Data entries in an index, 276 Data independence, 9, 15, 74:3 distributed, 736 logical, 15, 87, 7:36 physical, 15, 736 Data integration, 995 Data fvlanipulatioll Language, 16 Data ivianiplliation Language (D1vlL), 131 Data mining, 7, 849, 889 Data model, 10 multidimensional, 849 sernantic, 10, 27 Data partitioning, 7:30 skew, 7:30 Data reduction, 747 Data skew, 7:30, 73:3 Data source, 195 Data streams, 916 Data striping in RAID, 309 -:310 Data sllblanguage, 16 Data warehouse, 7, 678, 754, 848, 870871 SUBJECT 1056 dean, 871 extract, 870 load, 871 metadata, 872 purge, 871 refresh 871 transform, 871 Database administrators, 21-·22 Database architecture Client-Server VS Collaborating Servers, 738 Database design conceptual design, 13, 27 for an ORDBivIS, 79~J for OLAP, 85~3 impact of concurrent access, 678 normal forms, 615 null values, 608 physical, 291 physical design, 14, 28, 650 requirements analysis step, 26 role of expected workload, 650 role of inclusion dependencies, 639 schema refinement, 28, 605 tools, 27 tuning, 22, 28, 650, 667, 670 Database management system, Database tuning, 22, 28, 650, 652, 667 Databa ,es, Dataflow for parallelism, 7~31, 733 Dataguides, 959 Datalog, 818-· ·819, 822 aggregation, 8:31 comparison with relational algebra, 830 input and output, 822 least fixpoint, 825-826 lea '3t rnodel, 824, 826 model, 82:3 rnultiset generation, 8:32 negation, 827·828 range-restriction and negation, 828 rules, 819 safety and range-restriction, 826 stratification, 829 DataSpace, lOCH Dates and times in SQL, 140 DB2 Index Advisor, 665 DBA 22 D BI Ii brary, 2.52 DBMS DBrviS architecture, 19 DBl\IS vs as :322 INDF~ Disjunctive selection condition, DDL,12 445 Disk array, ~309 Deadlines hard VS soft, 994 Deadlock, 5:.n detection, 556 distributed, 756 global VS local, 756 phantom, 757 prevention, 558 Decision support, 847 Decision trees, 906 pruning, 907 splitting attributes, 907 Decompositions, 609 dependency- preservation, 621 horizontal, 674 in the absence of redundancy, 674 into 3NF, 625 into BCNF, 622 lossless-join, 619 Decorrelation, 506 Decryption, 709 Deductions, 820 Deductive databases, 820 aggregation, 831 fixpoint semantics, 824 least fixpoint, 826 least model, 826 least model semantics, 82:3 :Nlagic sets rewriting, 838 negation, 827·-828 optimization, 834 repeated inferences, 834 Seminaive evaluation, 836 unnecessary inferences, 834 Deep equality, 790 Denormalization, 652, 6E)9, 672 Dependency-preserving decomposition, 621 Dependent attribute, 904 DES, 710 Deskstar disk, 308 DEVise, 1001 Difference operation, 105, 141 Digital Libraries project, 997 Digital signatures, 71:3 Dimensions, 849 Directory of pages, :326 of slots, :329 Directory doubling, :175 Dirty bit, :.H8 Dirty page table, 585, 589 Dirty read, 526 Discretionary access control, 695 Disk spa.ce manager, 21, :304, 316 Disk tracks, 30ti Disks, :305 access times, 284, 308 blocks, ;306 controller, 307 cylinders, tracks, sectors, :306 head, 307 physical structure, 306 platters, 306 Distance function, 911 Distinct type in SQL, 167 Distributed data independence, 736, 743 Distributed databases, 726 catalogs, 741 commit protocols, 758 concurrency control, 755 data independence, 743 deadlock, 756 fragmentation, 739 global object names, 742 heterogeneous, 737 join, 745 lock management, 755 naming, 741 optimizatioIl, 749 project, 744 query processing, 743 recovery, 755, 7,58 replication, 741 scan, 744 select, 744 Semijoin and Bloomjoin, 747 synchronous vs asynchronous replication, 750 transaction management, 755 transparency, 7:36 updates, 750 Distri buted deadlock, 756 Distributed query processing, 743 Distributed transaction rnanagement I 755 Distributed transactions, 73G Division, 109 in SQL, 150 Division operation, 109 Dl\fL, 16 Document type declarations (DTDs),2:31 Docurncnt vector, 9:30 DoD security levels, 708 Domain, 29, 59 SU:BJECJ INDEX Domain constraints, 29, Gl, 7:3, 166 Domain relational calculus, 122 Domain-key normal form, 648 Double buffering,